4GHz CPU Battle: AMD 2nd-Gen Ryzen vs. Intel 8th-Gen Core

[Julio Franco]

We can say upfront that this article is in no way buying advice, but we're testing purely for the science of it. For the unaware, IPC (instructions per cycle) provides a good indicator of how fast a processor is, so with that in mind we're putting Intel and AMD's latest CPU architectures to the test running at a flat 4GHz.

Read the full article here.

https://www.techspot.com/article/1616-4ghz-ryzen-2nd-gen-vs-core-8th-gen/

 

[Puiu]

This turned out more of an benchmark of the cache than the IPC itself. The results vary a lot more than I expected. Nonetheless this should put the "speculations" to rest and give us proper results to work with.

 

[Steve]

This turned out more of an benchmark of the cache than the IPC itself. The results vary a lot more than I expected. Nonetheless this should put the "speculations" to rest and give us proper results to work with.

Isn't that all part of it?

 

[amghwk]

Sigh... we keep seeing these again and again, year after year... that AMD still trying to catch up with Intel, but could never really trounce it once and for all and claim clear superiority.

But, if AMD can slash at least 30% off it's current price of the CPUs, I think many will have second thoughts buying an Intel.

 

[Puiu]

Sigh... we keep seeing these again and again, year after year... that AMD still trying to catch up with Intel, but could never really trounce it once and for all and claim clear superiority.

But, if AMD can slash at least 30% off it's current price of the CPUs, I think many will have second thoughts buying an Intel.

If we're talking purely on gaming results then you might be on to something (when looking only at the 8700k and the 8600k), but AMD still holds a very decent lead in many of the productivity tests, something which someone like me can use (not needing to buy an expensive workstation class system for small to medium-ish tasks is godsent). Reducing the time it takes to make application builds is very important to me.
Pricing something purely on just gaming results is not smart business and you are even asking for a huge 30% decrease in price even though you already save a bit of money --> that's just you being entirely unrealistic and you clearly have a secret agenda.

 

[R00sT3R]

As someone who needs all the single core IPC he can get (flight simulators & emulation) these results look good for AMD at 4Ghz, but then I remember with an 8600/8700 I could very easily add a 15-20% overclock and leave the Ryzen's for dust in the area of performance I need the most.

..until Zen can match, or very nearly match Intel clock speeds, I think intel will still be my go-to CPU choice.

 

[Burty117]

Pricing something purely on just gaming results is not smart business and you are even asking for a huge 30% decrease in price even though you already save a bit of money --> that's just you being entirely unrealistic and you clearly have a secret agenda.

It's called "Fanboyism". I tend to filter out those comments* these days, kinda like how I scroll past adverts without giving them a seconds thought.

*With the exception of one Techspot member because he's hilarious.

 

[ET3D]

Thanks for the interesting benchmark.

 

[Puiu]

As someone who needs all the single core IPC he can get (flight simulators & emulation) these results look good for AMD at 4Ghz, but then I remember with an 8600/8700 I could very easily add a 15-20% overclock and leave the Ryzen's for dust in the area of performance I need the most.

..until Zen can match, or very nearly match Intel clock speeds, I think intel will still be my go-to CPU choice.

Yeah, I heard about flight sims only using a few cores (although a few of them did get some multithreading related patches), especially those that use OpenGL for rendering , but I don't know much about emulation beyond what I read about console emulation (which recently started moving towards much better multithreading, CEMU being an example). Are you talking about another kind of emulation?

 

[trparky]

Intel's superiority comes from the fact that you can overclock most of their chips like a mofo. With good cooling one can easily push 4.7 to 4.8 GHz on an 8700K and thus positively smoke the Ryzen 2600X which is stuck below 4.3 GHz. AMD desperately needs Ryzen to clock higher, much higher than it can right now if they have any hope of catching up to Intel because right now Intel is the performance king in gaming because of its ability to clock high.

 

[CBTex]

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

 

[trparky]

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

But if you were in market today for good gaming performance, Intel is your only choice.

 

[Techguy10]

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

Its down to the node they are currently on. The next process they shift to, for Zen 2 Ryzen CPU's will allow them to clock even higher. We're talking a LOT higher. The next node is a high performance node, for mobile and desktop. Those behind this did it different this time. If they clock as high as they expect, you can guarantee this conversation will be different next time. The people saying things about the 8700k may not be wrong: its a good chip, for today.

There's a window due to Intel struggling/having struggled to go beyond their current size, and though the area's are much the same when the two (future) processes are considered, the AMD side of manufacturing - like TSMC (Taiwan Semi conductors) and GloFo, are seemingly [we don't know exactly where Intel are at but its behind;] are far ahead. Pretty sure even Samsung is in there too. And if they move to 5nm before Intel too, well Intel have a lot of R&D money but it remains they are behind there.

This means in the coming years, by or around 2020/21 expect Ryzen Chips to be even better value. You may even be blown away. I use terms like maybe in spite of it being almost certain. AMD KNOW they have the architecture, and now the node.

We're talking way better, however and the fact that AMD will not get hammered on the yield [+ by the time Intel reaches that point] means more money in your pocket. Zen 2 is not being produced the same way as Zen 1. Its different silicon, on a different process, somewhere else per my latest information.

Smaller processes like 14/12/10, etc, get more expensive and the wiggle room to save some money as time goes on shrinks. This is another tac in the toe of Intel, to bring production costs down has been tricky. I don't think exactly what they're doing with price is highway robbery, but its convenient.

The actual cost of the chip is down to the increased yield's they can get (and some headroom we all know Intel have slapped onto people, but its not /too/ much), because AMD can cut out their CPU's on the wafers more efficiently, meaning they get more per wafer - and of course as we know then glue the 2 modules together. They ignore/discard the defective spots, and where possible they simply disable parts of the module. So when they wack the modules together they get a complete Ryzen CPU consisting of all the right parts inside. Then they slap on the better heat-spreader with proper solder around the thin silicon and are able to sell them for less.

Because Zen 2-Ryzen is going to clock higher, and because its a new architecture and not just a refresh, even with the latency issue, going forward there's a lot of excitement around the Zen 2 for those who know. Expect it to match or wizz past the intel chip

Its part conditions out in the market/industry and part because Zen 2 will clock higher + new improved architecture on a different node. There's a 30-40% power/performance margin to play with on the new node. The architecture/design is basically done.

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

But if you were in market today for good gaming performance, Intel is your only choice.

I know of what you speak, and technically in a first past the post scenario of 'engines' for games, you would be right. And I can't say you are wrong.

But whats missed in a lot of reviews as well is that it takes a 1080TI and a certain resolution to make the gaps even wider.

IF you did game on a 1080 however, or don't worry even a 1080TI and you used a Ryzen 2 refresh (the 2600x for example), or a ryzen 5 even, you would not be missing out on much.

In fact, nothing. This is why these things are flying out the door. And the conditions for that gap to be extended out in the environment/industry's no longer exist. They are not making any titles which will send the Ryzen scores backwards, and patches exist.

So while you are write, the actual scenario is more complex because systems and uses are so complex.

 

[Techguy10]

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

In addition to the TLDR but certainly interesting insight in my post above, where AMD is expected to trounce intel, now AMD know they not only the architecture but the node advantage, we see Intel's problems are manifold.

And like I say for the market with what I go into, it means they won't be as competitive as AMD for some time. However, the caveat is Intel has a wider array of area's they are into. Don't expect Intel to sink maybe or not make quality, competitive CPU's, but for a time, given the breadth of the situation, I could almost guarantee AMD will be smashing them in CPU price/performance before long. Of course we don't know the exact Zen 2 situation in terms of specs, even if I did, couldn't say, but Zen 2 seems like the issues discussed in the article simply won't matter.

I liked the article and its good that it was written but things change. IPC won't be anyone's concern come 2020. The entire disruption to the market would have taken place, and the price/performance/value of CPU's will be as thus. And the AMD's will be clocking a lot higher. I doubt as many people who bought intel this year will do so again in the early 2020's. This has been a remarkable turn around.

 

[Maxiking]

Yeah, Intel is doomed.

While they have already 10nm chips, AMD is selling Ryzen+ as 12nm and in fact, it is just on 14nm+ and they don't even compare those two processes but to 16nm finfet of TSMC.

Let's look at the old roadmap:

https://I.imgur.com/lhHiwIt.png

What is there? 14nm+, woah.

Do you remember the thing AMD has been spreading..

"12LP process offers a 10% performance improvement and a 15% circuit density improvement.."

But compared to what???

Look at this slide.

https://I.imgur.com/cQJnTPS.png


Woah, marketing is a thing.

 

[Techguy10]

Yeah, Intel is doomed.

While they have already 10nm chips, AMD is selling Ryzen+ as 12nm and in fact, it is just on 14nm+ and they don't even compare those two processes but to 16nm finfet of TSMC.

Let's look at the old roadmap:

https://I.imgur.com/lhHiwIt.png

What is there? 14nm+, woah.

Do you remember the thing AMD has been spreading..

"12LP process offers a 10% performance improvement and a 15% circuit density improvement.."

But compared to what???

Look at this slide.

https://I.imgur.com/cQJnTPS.png


Woah, marketing is a thing.

Lol yeah. I can only speak for myself and my posts. I'm trying to not make my posts too complicated, but it is true people say 12nm is comparable to 14 and the 10nm Intel will be on will be comparable to 7, in various ways, depending on what is being discussed. Its size, density, number of transistors, etc, or even which year we are talking about haha.

The foundary's in question have bought out IBM's assets, and a lot of work that was being done there has been rolled into the new process, and there's been a lot of progress on certain fronts, too in a short period of time. The last gen was typically (and unfortunately for desktop) in the grips of a mobile industry surge. But the desktop and server market has come back since those plans were underway. By the time we get to what is being called 7nm, there's actually a lot of extra steps to make a quality piece of silicon. It becomes exponential the steps involved to clean and process them - and for myself I am not talking about marketing slides and the reality of physics and the gates/transistors, and what certain measurements pertain to. There's different ways of measuring, so in actuality depending on who you're talking to, they'll either laugh at what you call 7nm or nod in agreement. And at the size of the past few generations of chips, since 14 lets frame it as, the difference is much, much smaller.

So in short: the marketing aspect you speak of has some legs and irony about it. However, I'm not speaking of those things, rather conditions; and if we're being honest, the last few generations you could get away with a bit of fluff talk.

^ That all changes though now they have moved to 7nm and are doing what they're doing on it. There's a kind of large jump in possibility when it comes to 7nm what they're doing with it; and another jump at 5nm.

After that, they're all staring a wall or looking to use a different material. So while its true it was all fluffy for some gens, there is a real and perceptible benefit to this node. And then next. The economies of scale have also lined up, so its a very exciting node to be on compared to the ones that preceded it. Because there is a legitimate increase in whats possible.

If Intel had of jumped to the next node by now, and not having competition meant they did not push (because also a node gets more cost effective as time goes on and crinkles in the manufacturing are ironed out, so why would they)... and of course no one can squeeze their head easily onto another/competitors node, so intel has intel.

To not digress further: Like any company, but from another angle is AMD, trying to boost their stocks. They need to investment dollars. Except when I look at all their marketing these days, and trust me I do go over it, I need to, its not a lot of fluff. The fluff ratio is waaaaay low on these Ryzens. Fluff factor- 5%? A few selective graphs - but they can back it all up.

____

The other reason I make mention of those things - and just talking about myself here... is of course by 2021, but probably 2020, is DDR5 Memory. Its all related to IPC and the Infinity Fabric. AMD will probably be first to DDR5 and we all know how the Ryzens like memory latency. Above that expect improvements around the Zen 2 design for this, and in anticipation of those things. These smaller components can be redesigned somewhat on a smaller node.

Basically when you tally up all the couple of per cent here and there, and over time, you arrive at a wolloping. Heck - if you run the current Ryzen+ with the 5% better memory configuration and a chip that clocks to 4.3 (though some go 4.4 and rare?4.5).. then you get to within 5% of the 8700k virtually always with IPC; and it certainly bests it on price/performance

I'm not saying Intel is going to die. I said their problems were manifold. And they are. I'm saying in relation to the IPC results we saw in these charts, its just all good news, and moreso going forward.

These things are already done, so marketing slides, and all of our 150+ cognitive biases, don't factor into it. They've already done that stuff. Intel is in a bind on the CPU front atm, wouldn't you be if you lost 20% market-share and were still trying to make the move to the next node make sense financially and technically?

Intel will be fine. I fully expect the 8700k to hold up very well going forward. It will slide by comparison to the newer chips, and don't expect the 8 core intel mainstream new one to clock very high, but they are in a bind; while on the graphics front you could say AMD is in a bind right now.

 

[Maxiking]

So maybe stop predicting the future, the same thing was being told about Nvidia in the past, how small AMD chips are the way to go and Nvidia is set to be unsuccessful with its big monolithic ones. How is the situation on the market?

After all the years they barelly matched what we have had here since Sandy Bridge. I will tell you a secret but Intel will introduce completely new architecture too on 10nm.

Let's wait how it turns out because I've read the same predictios in the past, Phenom, then Phenom II, then Phenom II X6, then Bulldozer. Some people never change.

 

[Techguy10]

So maybe stop predicting the future, the same thing was being told about Nvidia in the past, how small AMD chips are the way to go and Nvidia is set to be unsuccessful with its big monolithic ones. How is the situation on the market?

After all the years they barelly matched what we have had here since Sandy Bridge. I will tell you a secret but Intel will introduce completely new architecture too on 10nm.

Let's wait how it turns out because I've read the same predictios in the past, Phenom, then Phenom II, then Phenom II X6, then Bulldozer. Some people never change.

omg. Maybe you don't understand. These things are done, dude. I'm not predicting. The people who shook hands on DDR5 have done that too. All thats left is for time to pass.

Again, I am not predicting.

I do not intend to make this some kind of flame war. I know of intel 10nm. You didn't read my post. Remember the part about the window? Do you understand economies of scale?

And about the past, I addressed that too. Take some time to relax, because its complex of enough of a topic as it is.

I thought I framed it nicely when I said: AMD know they have the architecture and now they know they have the node too. Everything you need to extrapolate my meaning is within that sentence. They are also bringing more R&D in-house, like with the new memory stuff.

To date Intel have ripped the interconnect off their higher end chips and put cores together with that. I'm not saying they're not smart or they don't get whats going on; the difference with Ryzen is its new from the ground-up and they're pumping them out today. The new intel 10nm (on their own fabs), will be good; they have nullified the harsh meltdown/spectre issues....but the point is, don't expect a nice shiny new 10nm chip from them at a cheap price for the consumer the moment they do, because it doesn't work like that.

AFAIK they're doing 10nm on mobile for some time.

 

[Shumbo]

AMD seems to have solved half the equation in regards to catching Intel. They pretty much are equal in IPC. Hopefully now they can work on getting the frequencies up as well.

Either way, Zen has been great for the market. Competition is awesome.

But if you were in market today for good gaming performance, Intel is your only choice.

Don't be silly. Even a lower midrange Ryzen like the 1500x gives fantastic gaming performance. After you hit buttery smooth....it's the law of diminishing returns. Hell...an R3 1200 will run 9/10 games that exist on this planet like butter, if you don't have some fetish for needing to run every brand new game that is released.

 

[Techguy10]

Don't be silly. Even a lower midrange Ryzen like the 1500x gives fantastic gaming performance. After you hit buttery smooth....it's the law of diminishing returns. Hell...an R3 1200 will run 9/10 games that exist on this planet like butter, if you don't have some fetish for needing to run every brand new game that is released.

Further to what I was saying, intel will probably use 14NM for some time - and they'll be at 10nm for some things when others are at "7" - and they will all perform to varying degress because maybe intel has enough transistors on one node, etc: fit for purpose in relation to the marketplace; the thing that is lost on people who need the highest FPS somehow is that A CPU/node only needs to be 'so small' or 'so fast'.

---> So if Intel can be 'behind' on a node but still be fit for purpose -> Then why can't AMD CPU's be fit for purpose?

Why can it make sense for one and not the other. It just never occurs to some people. The graphs we see in the article, and in others, are demonstrable to that.

I cap my frames on an older Intel CPU at 80 frames! lol My card can go a lot faster but why bother.

The Ryzen CPU, does more than just push a lot of frames, the extra cores in a 2600x allow an entire computer to load balance better, and the CPU is fast enough to keep min frames from being low.

In my post above I did refer to Shumbo's post - but I was also seeing the posit differently. I don't entirely agree with the statement he made (but not in a bad way).

This then lead me to make a post about the behind the scenes stuff. And while its fair the other fellow who chimed in had some good points, whats lost in such discussions is that the Ryzen is well over the barometer... and lol... I don't care (or mind, like I could buy one, I don't know?) that Intel's next CPU is a different arch... what does it matter.... the Ryzen already exceeds the minimum requirements. And that situation will only improve.

Its an unintentionally false statement (via semantics) but not insincere, is all: AMD did solve the equation.

Thats what I was saying to the other fellow - expect IPC not to be of concern. As in, you won't dither either way over them. Even edge cases going forward won't stack up to your wallet. That doesn't mean Intel stands still, but it does mean AMD has caught up.

From my judgement the early 2020's, you won't see the purchasing landscape of the few years before 2018, when it comes to CPU's, rather an improvement on even this year.

 

[Techguy10]

Just for the TLDR crowd. The ELI 5 version with some more context. IPC is great on Intel and AMD at the moment in comparison to years gone by. Going forward IPC increase on AMD will be significant over previous chips, I'd wager heavily in comparison to Intel's going forward. But the Intel multi core jump is hitting a wall, new arch or not. Global Foundries for instance was one of the first iirc to put copper into CPU's. No one in the Founrdy business is slack behind the ears; but as you may assume correctly its a costly and large business to manufacture in since its a complex and exact process as far as manufacturing goes to make a CPU.

- However the Ryzen+ refresh did not leverage the entire benefits of the node- 12LP. So the refresh though it jumped to 12, did not utilize the full raft of things it could have on the 12 node. They took some feedback and incorporated it into things, but the long and the short of it, was they did not need to utilize the whole lot, and to keep costs down; because the node is ending and they're moving onto 7NM, so the lions-share of the man hours was moved over to the 7NM side. The year prior, the first Ryzens were on 14LPP. Coffee lake is also on a 14nm (nano meter) node.

- For the uninitiated, last years was on 14LPP. LP means low power, and LPP, low power plus. These were nodes made with mobile in mind. Despite this they do high performance well, but this is not what they were originally designed with to emphasize compared to the previous nodes. Global Foundries has the finfet technology, turning the gate on its edge basically, and its literally some of the best going tech and it was placed into their New York facility which at the time was brand new, so it just went in and that was that. But obviously there's a bunch of other considerations.

- On top of this we see a lot more work put into Zen 2. So Ryzen "3" will be on 7nm. On 7nm its a 'real' jump. They can make stuff twice as small almost, and have more room to play with how things are done. Its made to be, off the bat a high performance node in the truest sense of the word, going on what everyone has learnt from the 14LP/LPP days. The prior sized nodes, 14 and above were not 'real' jumps in terms of possibility. This one is, and there is a another pretty significant jump when they move onto 5Nm sized nodes.

- With the new node they expect to hit 5GHz or very near it. This is with some kind of power saving as well, which is needed.
- Then depending on how much it makes sense to make chips, since AMD is using a modular design to build CPU's with (Intel still have a monolithic/single large design), will depend on what core combinations we see, since this in turn affects how much money can be made from the materials they use, and the final price in store.

In other words, the IPC difference, is pretty much done. Intel is not making a new architecture specifically for speed enhancements.... the are making a new arch because the old arch was 'old'. Starting fresh will tie up loose ends, including the specter and meltdown.

But will it improve on IPC? I doubt it - not due to their ability, but the manufacturing, laws of nature, etc, etc. Expect IPC to be much closer from AMD going forward, and maybe expect it to surpass it.

On the multi-thread front, adding 2 cores to give us an 8 core intel in the mainstream, won't be the panacea people seem to think.

 

[Maxiking]

As I told with Nvidia vs AMD, they used the opposite approaches but used the same process and the same manufacturer. So without the knownledge how actually the future chips will perform this babble leads to nothing.

I don't have the insider info so I can't comment on anything because it would be just an uneducated guess. Also the fact Intel makes the chips on its own and AMD outsources them so their advantage of smaller, cheaper CCXs could be nullified that way. And that is all I can say for now.

 

[ZackL04]

Now if only these amd’s could near 5ghz like the Intels...

 

[Techguy10]

As I told with Nvidia vs AMD, they used the opposite approaches but used the same process and the same manufacturer. So without the knownledge how actually the future chips will perform this babble leads to nothing.

I don't have the insider info so I can't comment on anything because it would just an uneducated guess. You also ignore the fact Intel is making the chips on its own and AMD outsources them so their advantage of smaller, cheaper CCXs could be nullified that way. And that is all I can say for now.

Except its not babble. Where did I ignore the fact intel... etc? Thats what I am on about- they challenges there. Our trouble is you're just not on page 10 of the handbook like I am, you're not even at the contents yet.

About that you wrote: The discussion did not gloss for people. I think we're all literate enough to know where we're at. Its somewhat like you say, but just because you're certain you cannot be certain, does not mean its necessarily the case. We do know where its headed. When they sat down to organize 5 years ahead while they worked some people worked out the future node, and others worked out just what kind of chip to design, do you think they knew the future then?

The secret sauce is the Infinity Fabric. Its only a new occurance, so I don't expect it to factor into your future-think. Improvements and developments around that are in AMD's favor, as is scalability because its all modular and its happening right now. In terms of bringing that to market for Intel right now (I'm sure they are onto it) [but] its not there yet, it can't be done.

In the press they have people focussing on an 8 Core intel chip just at the moment, and timed certain announcements to coincide with announcements that made them look behind. Its about share price, dear fellow. They know how to inject into the media cycle.

Don't be so glad you think you can't predict things. The prices on that front/style for them just are not there at the moment. They know this - they don't have a modular design just laying about. Which is not to say they couldn't do one, but they were caught out.

__

-> So what I mean to say about the discussion up to this point - the article included - is that IPC is neither here nor there going forward. Forget about IPC - for the complexity of threads needed for IPC, given the limitations are approaching in terms of the current designs of CPU's... its mainly over as a race and Ryzen+ is within 5%, which to 98% of tasks is enough. With the improvements to the design of Zen 2 and other factors, of which more work is done on Zen 2 (so Ryzen 3 chips and beyond) IPC value to the consumer is not in question. Its barely in question except for edge cases right now as I stated. We're talking about consumer-grade chips, too.

Intel have muddled about with their interconnect. There's a reason Kaby Lake was done different to Coffee Lake. Thats the way things are headed.

Going forward its about more cores and price. IPC is at least at parity and with AMD increasing clock speed and with improvements to other factors in the design like power draw, it makes for a better chip. Its like this: wacking an extra 2 cores into an intel in the next few years won't be doing much overall.

Going forward, too, more things (but granted, not all) are going multi threaded. So the company's who bring affordable high powered multi-core CPU's to market are the ones who will stand to gain. In the marketplace we see the gain from AMD. This gain is almost certain to continue into the foreseeable future (no later than 2025 based on todays understanding).

IOW its not just about the classic CPU and IPC going forward. Its about higher core count and the ability to mass produce them to make a profit.

So we can indeed predict, or rather state, how things will be in the medium term. Improvements around the infinity fabric and other measures will bump up the IPC, and there's a new socket coming beyond Ryzen version 3 chips, and faster RAM. The 8 core intel chip having an effect on the market to change any of that is nil, as an example.

 

[Techguy10]

Now if only these amd’s could near 5ghz like the Intels...

next year or two. At or near enough. Not all Intel's do 5Ghz easily either btw. Some do. Some golden ones do 5.2. Many do only 4.8/9. There's a reason the 8700 k has an official base clock of 3.7 and a boost of 4.7....

plus to actually do that, you need solid cooling, a good board, luck, and it really helps to delid it; plus, either way the temps will be high and the power draw will be out of this world. Which Intel could give a bump to things if they did use a proper solder....

So when people say intel do 5ghz, they really 'don't' as such, but they do if you're prepared to blow its every metric out the water. Its not an efficient or cost effective 5ghz. If it was, they'd boost to that out the box.

But in terms of price/performance when AMD does to at or near 5Ghz, it will be with lower power draw on 7nm.